—— Ai ——
AI Core Concepts, Visually Explained
A quick visual tour — hover any box to see what it is.
The landscape
Large language models — a kind of generative machine learning — have become our front door to the wider landscape of “AI.”
Basic App Stack
Every AI app — Claude, Notion AI, your internal assistant — is the same three layers: App Logic on top, a Model in the middle, and Compute underneath.
More Sophisticated Apps
Apps grow more capable in two ways: running multiple models inside one App, or connecting one App outward to other apps, data, and software.
Multiple models, one App. One App Logic layer can orchestrate several model calls — a fast model for a first pass, a stronger one to refine.
One App, connected outward. An App can call another AI App, query a data source, or trigger traditional software through a connector like MCP or A2A.
Agents
An agent is the same LLM placed in a loop with a goal — think, act, observe, repeat — until the goal is met.
The model harness
One way to use these ideas: a harness coordinates several specialized model calls — plan, execute, judge — and loops until the output meets the bar.
Where it runs
Any model runs in one of two homes — cloud-hosted or self-hosted — and the choice decides who sees your data.
Inference providers
If you go cloud, three kinds of provider can serve the model: the model lab itself, a cloud platform, or a dedicated inference provider.